Transformers are often used to transfer information or power between circuits that are operating at different voltages or under different noise conditions. In many circuit arrangements, a logic signal must be transmitted between two circuits that must otherwise be electrically isolated from one another. For example, the transmitting circuit could utilize high internal voltages that would present a hazard to the receiving circuit or individuals in contact with that circuit. In the more general case, the isolating circuit must provide both voltage and noise isolation across an insulating barrier.
One type of galvanic isolator utilizes a transformer based system to isolate the two circuits. The sending circuit is connected to the primary coil of the transformer and the receiving circuit is connected to the secondary coil. The information is transferred by modulating the magnetic field generated in the primary coil. In this arrangement, the sending and receiving circuits can utilize entirely different power supplies and grounds and operate at different signal voltage levels. Typically, the transmitter and the two windings are constructed on a first semiconductor chip and the receiver is constructed on a separate chip that is connected to the first chip by wire bonds or the like. The two transformer windings are, typically, deposited over or near the drive circuits on the first chip by patterning two of the metal layers that are typically provided in conventional semiconductor fabrication processes. Alternatively, the coils may be fabricated on a different chip.
If the transformer coils are fabricated on the transmitter chip, the size of the transmitter chip is set by the size of the transformer coils, which typically require a significant area of silicon compared to the drive circuitry. Alternatively, if the coils are fabricated on the receiver chip or a separate chip, the coils will still require a significant area of silicon on those chips. The cost of the semiconductor substrate is a significant fraction of the cost of the isolator. This is a particularly significant problem when large coils are required to provide the coupling between the transmitter and receiver. In addition, many applications require multiple independent galvanic isolators on a single substrate. Cross-talk between the isolators constructed on silicon substrates using conventional semiconductor fabrication techniques is difficult to block in a cost-effective manner because of fringe fields generated by one coil being coupled to an adjacent coil. If the chips are separated by a sufficient distance on the silicon substrate, the cost of the wasted silicon becomes significant.
In addition to the wasted silicon area, devices constructed using conventional silicon integrated circuit fabrication have limitations that are imposed by the design rules of the fabrication line and the limitations as to materials that are allowed on that line. For many applications, the dielectric insulation between the coils of the transformer must withstand voltages in excess of 1000 volts. The thickness of dielectric that is available in conventional CMOS fabrication lines is insufficient to provide this degree of insulation. In addition, in some applications it would be advantageous to provide a ferrite layer or layers near the coils of the transformer to improve the coupling efficiency. However, the materials in question cannot be utilized in many conventional fabrication lines.
In some cases, it would be advantageous to power one of the circuits from the other circuit. For example, the transmitting circuit could power the receiving circuit. Such an arrangement would allow the receiving circuit to operate at different voltages than the transmitting circuit without requiring a separate power source on the receiving circuit. In principle, a transformer could also be utilized to provide the power transfer function. However, the efficiency required to provide the power transfer function is significantly greater than that needed to merely transmit information. Hence, such transformers are not easily, or economically, constructed using silicon-based fabrication techniques.
Miniature transformers constructed by winding wire around small cores are also known to the art. However, these devices are made one at a time, and hence, lack the economies of scale that are provided by wafer-scale photolithographic techniques and other mass production techniques developed for integrated circuits and the packaging thereof. Miniature transformers made by plating the coil pattern for the primary coil winding on one side of a printed circuit board and the secondary winding on the other side of the printed circuit board are also known. However, these dielectric core transformers have insufficient windings and are required to operate at relatively high frequencies because of the lack of a soft ferrite core.